A central issue in prostate cancer (PCa) is to recognize potentially lethal cancer at diagnosis, and to identify the causes and underlying mechanisms that distinguish lethal from indolent disease. Using a case-only design, we will develop a molecular signature for potentially lethal PCa by comparing the RNA expression profiles of tumor tissue from subsequently lethal PCa cases to tumor tissue from men without known lethal disease. Collaborating with researchers at the Broad Institute of Harvard University/MIT, we propose to apply a novel but proven high-throughput profiling technology to assess RNA expression in archival tumor tissue using a 24,000 gene platform. Our previous work provided converging evidence of a key role of the insulin-like growth factor (IGF) system in PCa risk and progression. We have already assembled an extensive prospective clinical and serological database on PCa with up to 26 years of follow-up. Germline polymorphisms and plasma levels of the IGF axis have been assayed in many cases who provided a prediagnostic blood sample. We now propose to extend this work with additional IGF/insulin components, including germline variations and tumor expression, in relation to PCa progression and mortality. Using a case-only design, we will assess circulating biomarkers and tagging germline polymorphisms in the IGF/insulin axis, comparing lethal cases to men without known lethal disease. We will also assess alterations of IGF/insulin signaling in PCa tissue (RNA and protein expression) in relation to fatal PCa, and will integrate plasma and genetic biomarkers with tumor tissue data to illuminate gene pathways that are dysregulated. Based on intriguing preliminary data, we will characterize tumor samples for presence of the common gene translocation, the TMPRSS2:ERG fusion, and address whether fusion positive tumors are more likely to progress when exposed to high levels of IGF/insulin signaling. The research will be conducted in the Physicians' Health Study (PHS) and Health Professionals Follow-up Study (HPFS) cohorts among incident PCa cases diagnosed from 1982-2008. We have assembled a PCa tumor repository of 1,600 cases (178 fatal) for tissue marker assays and are constructing high-density tissue microarrays for quantitative immunohistochemistry and FISH assays. We will have levels of circulating biomarkers measured in plasma (N=1,881) and SNPs assayed on extracted DNA (N=2,281). All men with PCa are followed intensively for information on treatment, PSA rise, metastases and cause of death with complete follow-up through 2012. The RNA expression data will greatly enhance our understanding of the influence of the IGF/insulin dependent and independent pathways on development of lethal PCa, which will aid in designing targeted therapy and prevention strategies. Most studies of PCa progression are based on elevations of PSA levels. A major strength of our proposal is that we use the most clinically relevant endpoint, lethal PCa. From the molecular signature of lethal PCa, we can identify a small number of highly predictive markers that could be ultimately assessed in biopsy samples. Thus, the findings can be translated to clinical practice, enabling clinicians to identify with confidence which tumors require aggressive therapy. Use of this rich resource of existing data and infrastructure permits a highly cost-efficient study, and the cross-disciplinary team of collaborators with a longstanding record of working together will ensure success of this project.